Forge Analytics and Reporting

A useful marketing report tells the reader what was measured, what was not measured and which decision the evidence can support. Forge's Performance area can help the team inspect supported-post metrics in the context of its Nodes and publishing surfaces. The reporting discipline comes from keeping definitions, observation windows and business outcomes explicit.

This guide focuses on the operational meaning of the numbers. It does not treat publishing access as proof of complete analytics coverage, or engagement as a substitute for verified commercial results. The Forge User Guide provides the documented metric and account-support reference used here.

Establish the reporting boundary

Before comparing results, list the connected accounts, relevant surfaces and metrics available for each. The documented integration supports available analytics for LinkedIn Company Pages, X, Facebook, Instagram and Threads, subject to platform permissions and format. LinkedIn personal-profile post analytics are not currently available inside Forge.

A successful personal-profile publication can therefore have no corresponding Forge analytics. That is a coverage limitation, not evidence of failure or zero response. If the team checks a platform directly, record that as a separate source with its observation date and definition.

Do the same for website and commercial data. Do not assume that the presence of a website article workflow means Forge collects all website sessions, enquiries or sales. Establish the actual measurement system for each outcome and verify any proposed handoff. The CRM integration guide explains how to assess such a workflow without assuming a native connector.

Use a metric dictionary before using a ranking

Metric	Working meaning	Interpretation limit
Impressions	Times the post was displayed	Repeated displays can involve the same person
Reach	Unique people or accounts reached, as reported by the platform	Definitions and deduplication belong to the source platform
Likes or reactions	Recorded reactions to the post	They do not identify buying intent on their own
Comments	Recorded replies	Count and commercial relevance are different questions
Shares	Reposts or equivalent sharing actions	A share is not necessarily an endorsement or a new unique reader
Clicks	Available link-click information for a relevant post	Clicks do not establish a completed enquiry or purchase
Engagement rate	The engagement measure shown for that platform	Compare only when the underlying definitions are compatible

Keep the source label where it matters. If one platform includes several interaction types in its engagement measure and another uses a different definition, the percentages are not automatically comparable. A report should explain the difference or keep the results separate.

Do not add reach across posts and call the total unique campaign audience unless the source provides the required deduplication. The same person may have seen several publications. Similarly, adding clicks produces recorded click events, not necessarily distinct visitors.

Give every observation a time basis

Metrics arrive after publication. The guide indicates that first measurements for supported posts are expected around 6 hours after publication, with platform availability and permissions affecting timing. It describes more frequent refreshes early in a post's life and less frequent refreshes as the post ages.

For a report, record when the post was published, when you observed the metrics and the age of the post at observation. A post measured after 7 days has had a different opportunity to accumulate response from one measured after a few hours. Comparing them without acknowledging that difference can mislead the review.

Choose a consistent observation window where practical. For example, a team can compare its selected posts after 7 days, while keeping a separate operational check for newly published content. The choice is a reporting convention, not a guarantee that all audience response happens within that window.

For older posts, use the latest information actually shown and seek Support if a refresh is unclear. Do not invent a refresh schedule for an interval that the current guidance does not specify. A static older value and a live recent value need different interpretation.

Treat missing values as a finding

A dash can mean that data has not arrived or is unavailable for the account or format. Leave it marked as missing until you establish the cause. Entering zero changes the meaning of the report and can make a supported platform appear to outperform an unmeasured one.

Use a brief missing-data note with the affected account, metric, post age and connection status. Then classify what you know: pending collection, unsupported coverage, a connection issue or an unresolved cause. These are reporting categories for the team, not named product statuses.

When missing data affects a calculation, state the coverage. “Rate calculated from the 2 posts with both impressions and clicks available” is clearer than a campaign percentage that silently mixes complete and incomplete records. If the missing posts are materially different, even the complete subset may not represent the campaign.

Calculate rates with matching numerators and denominators

The following hypothetical report uses 3 posts from the same account, each inspected after 7 days. It demonstrates arithmetic, not Forge customer performance. The last post has a click value but no available impression value at the time of review.

Post	Impressions	Link clicks	Clicks ÷ impressions
A	700	14	2.0%
B	1100	11	1.0%
C	Unavailable	5	Not calculated

For the complete records A and B, the aggregate rate is 25 ÷ 1800, approximately 1.39%. Averaging the two percentages would produce 1.5%, which gives the two posts equal weight despite their different impression counts. Use the aggregation that answers the question you have stated.

The report can show 30 recorded clicks across all 3 posts, but it must not divide 30 by 1800 and call that the campaign click-to-impression rate. That numerator includes Post C while the denominator excludes its unavailable impressions. Keep the incomplete post separate or wait for the missing information.

This custom calculation should be labelled “clicks ÷ impressions”, not automatically renamed “engagement rate”. The platform's engagement metric may include other actions or use another denominator.

Connect the numbers to message context

Forge's retained context makes it useful to examine which Node, buyer, pain, proof and surface were associated with a publication. A ranking can identify where to investigate. It does not explain the result by itself.

Suppose two posts address the same buyer but use different proof. One presents a clear example of a deliverable; the other describes a broad benefit. If the first produces more relevant questions, the team may have a reason to explore that evidence-led approach. It still needs to examine distribution, timing, wording and the actual questions before making a strong claim.

Write observations and interpretations separately. “Three comments asked about eligibility” is an observation if the comments support it. “The market prefers our new positioning” is a much broader interpretation. The Performance Loop lesson shows how to turn a narrow finding into a justified input update.

Keep attribution and commercial outcomes in their proper place

Attribution assigns credit for meaningful actions across touchpoints according to an attribution model. That is the role described in Google Analytics' attribution guidance. It is a different question from which message components were present in a popular post, and it is not established merely by retaining marketing context.

If the business reports enquiries, qualified opportunities or sales, identify the system and definition for each. State whether a source was tracked, self-reported or unknown. Keep a content click separate from a website session, and a submitted form separate from a qualified sales conversation.

Do not present a pre/post improvement as proof that Forge caused incremental revenue. Other activity, demand changes, distribution and measurement changes may be involved. For the conceptual distinction, read Marketing Attribution vs Marketing Intelligence.

Make the weekly report short enough to act on

A decision-ready report can contain a delivery summary, measurement coverage, a few comparable observations, relevant business outcomes from their verified sources and the decisions proposed. Link to the detailed records rather than copying every metric into the meeting document.

• Delivery: what was intended and what publication was confirmed.
• Coverage: which accounts, posts and metrics are included or missing.
• Observations: what changed within stated definitions and windows.
• Interpretation: plausible explanations and remaining uncertainty.
• Action: the input, publishing or follow-up decision and its owner.

Include operational quality. A campaign with attractive engagement but repeated wrong-account incidents needs a publishing correction. A strong message with no available next step needs an offer handoff correction. The report should reveal those issues rather than hiding them behind a single score.

Use Campaign Review for a broader retrospective and Multi-Channel Publishing when coverage differences reflect different account or surface choices.

Ask for the measurements your decision needs

Before adopting a reporting routine, test your actual accounts. Confirm the available metrics, the initial collection experience and how missing values are shown. Bring the commercial outcomes you need to measure and identify which belong in other systems.

Request a Forge demo focused on analytics coverage and your reporting questions. A useful session should establish what your team can observe, what it must verify elsewhere and how those records support a clear decision. The Forge overview explains the platform's wider role in the marketing workflow.